Dynamic requirements for a functional protein hinge

Abstract

The enzyme triosephosphate isomerase (TIM) is a model of catalytic efficiency. The 11 residue loop 6 at the TIM active site plays a major role in this enzymatic prowess. The loop moves between open and closed states, which facilitate substrate access and catalysis, respectively. The N and C-terminal hinges of loop 6 control this motion. Here, we detail flexibility requirements for hinges in a comparative solution NMR study of wild-type (WT) TIM and a quintuple mutant (PGG/GGG). The latter contained glycine substitutions in the N-terminal hinge at Val167 and Trp168, which follow the essential Pro166, and in the C-terminal hinge at Lys174, Thr175, and Ala176. Previous work demonstrated that PGG/GGG has a tenfold higher Km value and 10(3)-fold reduced k(cat) relative to WT with either d-glyceraldehyde 3-phosphate or dihyrdroxyacetone phosphate as substrate. Our NMR results explain this in terms of altered loop-6 dynamics in PGG/GGG. In the mutant, loop 6 exhibits conformational heterogeneity with corresponding motional rates <750 s(-1) that are an order of magnitude slower than the natural WT loop 6 motion. At the same time, nanosecond timescale motions of loop 6 are greatly enhanced in the mutant relative to WT. These differences from WT behavior occur in both apo PGG/GGG and in the form bound to the reaction-intermediate analog, 2-phosphoglycolate (2-PGA). In addition, as indicated by 1H, 15N and 13CO chemical-shifts, the glycine substitutions diminished the enzyme's response to ligand, and induced structural perturbations in apo and 2-PGA-bound forms of TIM that are atypical of WT. These data show that PGG/GGG exists in multiple conformations that are not fully competent for ligand binding or catalysis. These experiments elucidate an important principle of catalytic hinge design in proteins: structural rigidity is essential for focused motional freedom of active-site loops.

Figure 1

(A) cTIM dimer (open, 1TIM25) with color coding of selected secondary structure elements. Residues participating in the active site are colored red, and include Asn11, Lys13, His95, Glu97, Glu165, Gly171, Ser211, Gly232 and Gly233. Among these, directly relevant side chains are shown for Asn11, Lys13, His95, Glu97 and Glu165. Active-site coloration trumps colors of other structural elements. (B) Close-up view with labeled active-site residues and additional coloration of N- and C- terminal hinges. (C) Active site in closed, liganded structure (1TPH26). Structural representations here and in Figures 2 and 4 were prepared using PyMOL, and PDB coordinates 1TIM and 1TPH for apo and bound forms of cTIM.

Figure 2

Various views of the packing of loop-6 hinge side chains for the apo and bound forms of WT chicken TIM. (A)–(D) Packing of the C-terminal hinge (Lys174, Thr175, Ala176), with hinge side chains represented by translucent space-filling yellow spheres, while solid spheres represent atoms of surrounding residues. (E)–(F) Packing of the N-terminal hinge (Pro166, Val167, Trp168), with translucent purple spheres for hinge side chains and solid spheres for surrounding groups. In all parts, blue spheres belong to central residues of loop 6, while purple and yellow (solid or translucent) belong to its N- and C-terminal hinges, respectively. In (A)-(D), light pink are Asn216 and Leu220, hot pink is Ser211, lime green are Gln180 and Ala181, green is Tyr208 and dark blue is Thr177. In (E)-(H), cyan is Arg134, orange are Glu129 - Leu131, lime green is Gln180, green is Tyr208, maroon are Ser96 and His100 and red are Tyr164 and Glu165. In the bound forms (at right) the substrate is shown as in brown stick representation. For all parts, the light gray cartoon backbone is the monomer containing the highlighted loop-6 region, while the second monomer is light yellow.

Figure 3

¹H-¹⁵N HSQCs of (A) apo WT cTIM (blue, loop-6 sites in red) with overlaid 2-PGA-bound WT spectrum (green, open contours, loop 6 sites in magenta) and (B) apo and 2-PGA-bound PGG/GGG spectra with color and contour schemes as in (A). For WT spectra in (A), purple arrows highlight each loop-6 apo-to-bound chemical-shift trajectory, though within loop 6, apo-form Gly171 and bound-form Trp168 resonances have not been identified. Additional residues indicated include the catalytic residue, Glu165 (not identified in the bound form), and other active-site participants.

Figure 4

False-color representations of (¹H,¹⁵N) chemical-shift differences mapped onto the cTIM dimer. (A) Δ_HN WT (apo – 2-PGA-bound), (B) Δ_HN PGG/GGG (apo – 2-PGA-bound), (C) Δδ_HN (apo PGG/GGG – apo WT), and (D) Δ_HN (bound PGG/GGG – bound WT). Gray coloration indicates a residue unassigned in one or both of the compared forms. To facilitate comparison, apo-cTIM coordinates (1TIM25) were used in (A) –(D), regardless of binding state or mutation. Note, the scale in (B) is 5 times smaller than in (A), (C), or (D).

Figure 5

(A) 2-PGA-bound PGG/GGG mutant spectra of loop-6 residues: (top) ¹H-¹⁵N HSQC and (bottom) ¹H_N-¹³C HNCA spectral slices (blue) with HN(CO)CA slice overlay (gold), both at fixed ¹⁵N. Conformational heterogeneity is evidenced by multiple HSQC peaks per residue, assigned by the corresponding distribution of intensity in triple-resonance spectra. Gly171 and Ile170 show particularly striking heterogeneity, while Thr177, the +1 residue of the C-terminal hinge, is apparently uniform. Fixed ¹⁵N shifts for HNCA slices are at 123.4 ppm (Ala169), 119.7 ppm (Ile170), 113.4 ppm (Gly171), 113.0 ppm (Thr172, Thr177) and 111.2 ppm (Gly173). Not all resonances associated with a given site appear in the ¹H-¹³C strips at discrete ¹⁵N chemical shift. In (¹H,¹⁵N) HSQCs, resonances from sites outside of loop 6 are not labeled. (B) 2-PGA-bound wt cTIM spectra loop-6 glycine residues: (top) (¹H,¹⁵N) HSQC and (bottom) (¹H_N,¹³C) HN(CA)CB spectral slices at fixed ¹⁵N. Only an HN(CA)CB spectrum was collected for the bound WT, thus only glycines, which reveal their C_α resonance in this C_β-directed experiment, are shown to simplify comparison with (A). Fixed ¹⁵N shifts for HN(CA)CB slices are at 122.9 ppm (Gly171) and 112.1 ppm (Gly173).

Figure 6

R₂ values by residue at 14.1 T for (A) WT cTIM in apo and 2-PGA-bound forms, (B) apo and bound PGG/GGG cTIM, and (C) in and adjacent to loop 6 for apo and bound WT and PGG/GGG. (D) is as (C), but at 18.8 T and lacking bound-form PGG/GGG. Shaded areas are provided to highlight labeled elements.

Figure 7

J(ω_N) values by residue at 14.1 T for (A) WT cTIM in apo and 2-PGA-bound forms, (B) apo and bound PGG/GGG cTIM, and (C) in and adjacent to loop 6 for apo and bound WT and PGG/GGG. (D) is as (C), but at 18.8 T and lacking bound-form PGG/GGG. Shaded areas are provided to highlight labeled elements.

Scheme 1

TIM-catalyzed reaction scheme (bracketed) with enediolate transition state and adjacent structure of the reaction intermediate analog, 2-phosphoglycolate (2-PGA) at right.

Scheme 2

Highly conserved loop-6 sequence in TIM, here labeled for amino acids 166 – 176 found in chicken (Gallus gallus) TIM. Nomenclature used in the text assigns N1, N2, N3 to N-terminal hinge residues 166-168 and C1, C2, C3 to C-terminal hinge residues 174-176 respectively.

Scheme 3

Kinetic model for catalysis by the PGG/GGG cTIM mutant, including equilibria involving nonproductive conformations, (E)* and (E^#S), of the apo and substrate-bound enzyme.